What Will We Do When AI Can Do Everything?

What can we do that AI can’t do?

I’ve thought about this for years, ever since Big Blue beat Garry Kasparov in chess over 20 years ago. Now AI agents are replacing virtual assistants, lawyers, and all kinds of roles. A lot of people fall into one of two camps: those who believe robots will take over everything and we should just wait for universal basic income, and those in denial who insist AI will never touch their job.

I think the truth is somewhere in the middle.

AI is a coworker, not a replacement

AI is becoming more like a virtual coworker. It can handle a lot of the stuff we don’t want to do, the tasks that aren’t the highest-value use of our time. Editing a video, scanning documents, data entry. AI can do that better, and it should.

But here’s the thing. Just because you have a calculator doesn’t make you better at math. Just because you have a scalpel doesn’t make you a better surgeon. The tool amplifies the person using it.

My mentor, the former CEO of American Airlines, taught me this lesson 30 years ago. I told him the internet would kill travel because people would just meet over video calls. He laughed and said they were ordering more airplanes, not fewer. The internet actually encouraged more travel. AI works the same way. It doesn’t eliminate human value. It multiplies it.

Find your highest-value skill and multiply it

The real question is: what should you and I focus on?

Focus on things that create value on top of what AI can do. Where do you have the most powerful relationships? What’s the one skill you have that, if there were 10 of you doing it, would produce 10 times more output?

Take Marko Sipilä, who’s exceptional at Facebook ads for HVAC companies. He’s using AI to scale his agency and build software to help even more companies. AI is multiplying his value, not replacing it.

Or George Paladichuk, who built an AI voice call center that answers phones when humans don’t. That system drives half a million dollars a month in trackable revenue that wouldn’t have existed otherwise.

The reward loop is everything

The reason software engineering is being disrupted first by AI is because the reward loop is built in. Code either compiles and works, or it doesn’t. That makes it easy for AI to self-improve.

The next industry in line is home services. Why? Because you can look inside ServiceTitan, FieldRoutes, Jobber, CallRail, and tie revenue directly back to marketing activities. That creates a measurable reward loop.

When your AI agents have clear metrics to optimize against, they work like children earning points at a fair. They get better and better at driving the outcomes that matter to your business.

Turn your knowledge into AI-powered systems

If you have courses, SOPs, or documented processes, you already have the raw material. A course is just a recipe: initial ingredients, a set of steps using specific tools, and an end product.

When you feed that recipe to AI agents, courses become implementation programs. You can deploy a swarm of agents to handle digital marketing, operations, SEO, Google ads, reputation management, and more, all based on work you’ve already documented.

The agencies and businesses winning right now aren’t afraid of AI. They’re getting more work done in a more measurable way because they’ve built the systems that let AI optimize toward real business results.

The people who win with AI aren’t fighting it. They’re identifying their highest-value skills, documenting their processes, building reward loops, and letting AI handle the 80% so they can focus on the 20% that drives real impact.

If you had 10 times more impact and only worked 20% of the time, you’d still have double the output. That’s the math that matters.